Forecasting Power Consumption by Deep Learning Approaches in the State of Telangana

Electricity is a vital part of modern life and important for the world economy. India is the third-largest producer of electricity in the world. There is a need to conduct continuous research on forecasting power consumption to meet the demand for overall increased consumption in various sectors. This research will enable policymakers to plan. In need of this research, an analysis was done to forecast electricity consumption. We have used the traditional and newly developed advanced models to predict the accuracy values. A traditional model like ARIMA and distinguished techniques Long Short Term Memory (LSTM) and Convolution Neural Network (CNN) have been used to forecast power consumption demand with a certain degree of accuracy between the years 2014-22. The analysis fits well with the data for LSTM showing an optimal fit than ARIMA and CNN for the observed data considered over eight years in the Industrial-HT sector of power consumption in Telangana State Southern Power Distribution Company Limited.